Drive layout optimizations using WindData objects #822
Conversation
floris/tools/optimization/layout_optimization/layout_optimization_boundary_grid.py
Outdated
Show resolved
Hide resolved
floris/tools/optimization/layout_optimization/layout_optimization_pyoptsparse.py
Outdated
Show resolved
Hide resolved
floris/tools/optimization/layout_optimization/layout_optimization_base.py
Outdated
Show resolved
Hide resolved
|
I merged |
|
Merged v4 back in again |
|
|
||
| from ....logging_manager import LoggingManager | ||
|
|
||
|
|
||
| class LayoutOptimization(LoggingManager): | ||
| def __init__(self, fi, boundaries, min_dist=None, freq=None, enable_geometric_yaw=False): |
There was a problem hiding this comment.
Does this mean that, if the user would like to provide a weighting to different wind speed bins; they must provide a wind_data object (whereas if they want to use uniform weighting, just supplying fi will suffice)?
Could we instead leave thefreq option, which will be ignored if wind_data is None, but which otherwise will be used in the creation of the TimeSeries object in LayoutOptimization.__init__()?
There was a problem hiding this comment.
@misi9170 let's chat about this one tomorrow,
There was a problem hiding this comment.
Ok good point @misi9170 , as discussed changing wind_data to required input to avoid supplying a use case no one is likely to want but could get accidentally
Drive layout optimizations using WindData objects
This pull request reorients the layout optimization classes to accept WindData objects (WindRose or TimeSeries), in place of freq tables, in order to facilitate allowing change of optimization from maximizing AEP to maximizing other quantities such as revenue. Changing to passing the WindData objects facilitates this because these objects hold the value per bin (WindRose) or as a time series (TimeSeries).
(Note this will appear to be many changes at first since built off of #819 ) but once that PR is merged, this will look much simpler
Related issue
#779